Native implementation of the simulator

Anki FSRS
61

Thanks for the update and I appreciate your hard work. Hopefully, this feature will be considered for a future release.

65

I read above that you plan to place the simulator on separate single page, kindly use interface similar to this with these many option, using parameters box in place of starting ease and interval modifier. This would allow experimenting with different options and how they would impact workload. The options next to “Cards” below “simulation options” are really useful.

The separate page would allow the graph to utilize the bigger screen size making it a better visual experience.

1 Like
66

Unfortunately, I’m not good at frontend development. So I buried the plan. Maybe someone could help me implement it.

2 Likes
67

I’ve been trying out the simulator and I find it quite interesting. Thanks for developing it!

I think I would like one extra feature, and I wonder what you think of it.

I see there are

  • “Additional new cards to simulate”. I take this to mean “imagine you added n new cards right now”.
  • “New cards per day”. I take that to mean “how many new cards will you view per day”.

Instead of imagining that the new cards are all added right now, I would like to simulate adding them at a daily rate. So that would rename the current “new cards per day” to “new cards viewed per day” and create a new field “new cards created per day”.

The reason is that

  • Adding cards daily is what I typically do, rather than adding 6000 new cards at once, so it’s more similar to my behavior, and
  • I have a better understanding of how many cards I typically create per day, than I do of how many cards I’ll eventually have created, so I can better understand what numbers to put into the fields, and
  • Daily versus one-time new-card creation may have a different result in the simulation. If I only create a fixed number of cards, rather than creating new cards, I’ll eventually run out of those, so long-term projections will only include a fixed amount of cards, whereas I’m likely to have a continually increasing card count not reflected in the simulation.
    On the other hand, if my “new cards viewed per day” exceeds my “new cards created per day”, I might run out of new cards on some day, so I won’t hit my “new cards viewed per day” that day, because there aren’t any new cards to view. The next day I will add more new cards to view, and those would appear in the simulation for the next day. This “no new cards remaining” behavior isn’t captured in a simulation that has a fixed number of new cards, but would be captured by “new cards created daily”.

It would also be interesting to see, for each day, how many New cards have been created but not yet reviewed. I might want to know, for example, if I’m creating so many new cards that I’ll never get to them all.

I’m interested in your thoughts on this. Cheers!

2 Likes
68

It doesn’t really matter for the simulation though, does it?

Imagine this:
1. Scenario:

  • You add 100 cards today,
  • and learn 10 new cards per day for 10 days.

2. Scenario:

  • You add 10 cards every day,
  • and learn 10 new cards per day for 10 days.

The important calculation is amount of new cards / amount of days to learn them. In your case:
1. Scenario:

  • You simulate 10 days.
  • In these days 100 new cards will be added in total on day one.
  • Calculation: 100 new cards / 10 days = 10 cards per day.

2. Scenario:

  • You simulate 10 days.
  • Every day you add 10 cards (which is 10 cards * 10 days = 100 cards total).
  • Calculation: 100 new cards / 10 days = 10 cards per day.

In both scenarios the result is the same: 10 new cards are learned every day. Whether you added the cards in bulk or every day doesn’t make a difference.

3 Likes
69

Another feature idea.

When I have several simulations in the figure at once it’s hard to remember what they all are. It would be helpful to see the simulation parameters in the legend, or in a table underneath, with each row representing a simulation.

1 Like
70

I’m afraid that would be too cluttered

71

I’m not good at frontend design. Do you have any best practice we can follow?

72

An idea could be to restore the parameters to the ones used to create a line whenever that line is right clicked in the legend?

1 Like
73

Idk best practice, but what occurs to me is: a table underneath the graph with one row of parameter values per simulation, with a line-color column. For extra points, mouse-hovering over a graph line would highlight the corresponding table row, and hovering over a table row would highlight the corresponding graph line (possibly dimming the others).

74

this analysis is correct only if you add an amount of cards equal or greater than your new cards/day x number of days simulated. however for a person with 20 new card limit, 100 added immediately looks different to 10 added over 10 days.

1 Like